The previous poster is correct. And semaglutide alters signaling which was damaged or ineffective.

And I'm 100% for it. Lots of diseases have improper signaling and altering energy metabolism is probably clinically risky. The hunger signaling is so deep in evolutionary history it's not possible to override in practice. Changing the setting also is safest--hunger signals are still there with their normal function just lowered in intensity.

People with the problem feel like they are eating 1200 calories even though they aren't. That's a big problem that deserves to be fixed.

> So this all seams horrible but maybe we should ask some philosophical questions about how wrong it really is to offer healthy-bmi-the-easy-way as a subscription.

About as wrong as offering the healthy-blood-pressure-the-easy-way as a subscription. Or clean-water-on-top as a subscription.

And in truth, the collection of BP medications has saved many people from nasty diseases and poor health.

> In some people, all calories are stored as fat regardless of how little they eat.

That can't happen because there's continuous energy consumption from something. That the hunger signals are wrong and misguided in some people is a certainty though.

> two people eating the same diet may not experience the same changes to their body's mass.

thats the calories out

> All you'll do is make yourself addicted, dependent on expensive weight loss shots to make you feel a little bit better.

Does that apply to cholesterol and high blood pressure medications?

Realistically, they are much more efficacious than lifestyle changes, and most people have few serious side effects and it stacks upon personal behavior changes.

And healthy eating is much much easier according to those who are on the GLP-1As. Hunger drives are obviously strongly evolutionarily embedded to the beginning of life.

And the new class of drugs (unlike previous generations) are showing clear clinical benefit to heart diseases, liver, kidneys and reduce chance of Alzheimers. It's pretty remarkable and I'd feel a fuckload better not having a heart attack if I had a problem. I'm opposed to such sort of sneering moralizing bullshit like "All you'll do is make yourself addicted, dependent on expensive weight loss shots to make you feel a little bit better."

I'm addicted to indoor plumbing and safe water treatment to reduce infections and illness. I'm 100% pro civilization. GLP1As, like blood pressure meds, are part of the deal.

Reusing the orbital craft is much more difficult to do economically than the booster. It re-enters far faster with much more damage---or would take much more fuel to slow down, severely reducing paying payload.

I suspect SpaceX may go to an expendable upper stage (light and with higher capacity as it doesn't need to survive coming back) for numerous missions, particularly those which need extra orbital energy.

There was the beginning of scientific computing using GPUs, CUDA was under development and the trend towards general purpose two-way computation was starting.

That 23% was likely not curtailed, a large amount was probably exported.

The curtailment of wind and solar is pretty low now with the large amount of battery capacity.

Batteries in CAISO went from 500 MW in 2020 to about 11,000 in 2024. Yes, that's power and not energy, but most are sized at 4 hours.

https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.caiso.com%2Fdocument...

They aren't good at deeper reasoning but they're good at memorizing and doing simple applications. The problem there, as a smart search engine, is that they memorize insufficiently well and don't know when they're making mistakes.

Parrots and corvids can reason and solve novel problems.

It's totally the other way around. The board wanted to fire Sam Altman because he was a podcasting bro and was taking over (now successful) and removing all ethical constraints on OAI and generally being a giant rich douchebag and taking tons of stock for himself.

The problem was ever hiring him in the first place. If that hadn't happened, Ilya would be CTO, and someone decent would be CEO, and openAI would be open and changing the world for good.

What makes you think the top NVidia processors like the H100 and G200 are not custom hardware tuned for NN processing? If one started from scratch, where are the main inefficiencies that would be solved, and at what cost? Would one invent something substantially better than what NVidia has now? Its doubtful.

The main use case for NN training does require highly programmable and flexible parallel processing like graphics computations do. BTC mining is a single computation but NN work is not as simply stereotyped.

> I suspect it would be more fruitful to increase efforts on converting existing models to quantized integer models and just stick with the existing hardware. This also massively lowers the power consumption by replacing floating point instructions with integer ones.

This is already done. And it doesn't lower the power consumption that much.

It's useful for using already trained nets, but not for training where the dynamic range of floating point is essential. There is lots of existing work in the literature on compressing/sparsifying and quantizing an already trained net to make it more efficient in production inference/scoring with low loss of effectiveness. This is lossy compression.

There is already bfloat16 and float16 and maybe even float8 hardware support in NVidia processors anyway. This new one is being proposed as an additional one which could be even higher performance if supported directly in hardware.

> Instead of summing thousands of numbers in a vector, perhaps maybe we should break with tradition even further and go analog?

There have been attempts in this direction, implementing neural networks in analog electronic hardware, for literally decades and a number of failed startups.

I believe a major problem is fabrication differences: one chip is different from another and you can't replicate results. So a net you train on one chip will not work properly on another. That makes so much of a human burden (figuring out how to recalibrate if possible between hardware) that they are not feasible products.

If it's inference only, then its going into a production situation and needs to be cheap, but moreover in something like that (e.g. automotive) it needs to be mass producible and reliable and not require expert fiddling.

> "Your tires will wear out much faster due to the increased weight from that battery."
> That is a lie. BEV tires wear out faster because the the BEV market right now specs lower mileage, low profile performance tires.

What's really happening is that new car manufacturers are specifying tires with lower initial tread depth to improve efficiency. So the new car tires wear out quickly. This is a new car problem, but more significant in EVs as rolling resistance matters more for efficiency, so there is more motivation to do this.

When you get your next tires, they will last longer if you choose correctly.